agentleFS
infeniainfenia/yukta/core/CLAUDE.md

yukta / core

CLAUDE.md1 starsChanged 20 days agoApache-2.0View on GitHub

The file

We split this CLAUDE.md into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    All service logic uses Project Reactor (Mono, Flux).

    line 18
  2. Very specific:

    Use Java records for data models (preferred) or Lombok (@RequiredArgsConstructor, @Getter) for immutable objects.

    line 35
  3. Distinctive:

    Use Jakarta Bean Validation.

    line 39
  4. Very specific:

    Follow the tiered "Test Execution Policy (speed)" in the root CLAUDE.md: run --tests for the classes you changed first.

    line 70

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

Core Module CLAUDE.md

Overview

Core service layer and workflow orchestration engine.

Build & Development Commands

# Run tests for core module
yk gradle :core:test

# Run a single test class
yk gradle :core:test --tests com.infenia.yukta.service.orchestrator.WorkflowOrchestratorTest

Key Architecture & Patterns

Reactive Streams (Core Pattern)

All service logic uses Project Reactor (Mono, Flux).

  • Services use Mono for single operations, Flux for streams.
  • Avoid blocking operations.

Workflow Orchestration

  • com.infenia.yukta.service.orchestrator.WorkflowOrchestrator: DAG orchestration logic, split across assembly/, compiler/, preparator/, strategy/, stream/, tracker/, and validator/ subpackages.
  • tracker/ keeps the queue and the lookup maps in DefaultTaskTrackerService, one run's state in WorkflowState, the spans and meters in RunTelemetry, applying queued events in TrackerEventApplier, the live log and status streams in ExecutionStreams and the remembered finished runs in ExecutionHistory. DefaultControlBusGateway delegates starting (admission, shared prepare, idempotency) to WorkflowStarter. It also delegates node commands to NodeCommands, restarts to RestartCoordinator and read-only views to ExecutionQueries (ControlDispatch is their shared base).
  • com.infenia.yukta.service.plugin.PluginRegistry: Plugin registration.
  • Executes a Directed Acyclic Graph (DAG) of nodes and edges.

Immutability & Records

Use Java records for data models (preferred) or Lombok (@RequiredArgsConstructor, @Getter) for immutable objects.

Validation

  • Custom validators in com.infenia.yukta.validation.
  • Use Jakarta Bean Validation.

Plugin logging

  • A plugin writes live log lines through the PluginLogSink that the engine puts in the Reactor context (WorkflowContext); the lines go to TaskTrackerService.appendLog and from there to PluginLogStore (InMemoryPluginLogStore, fed by LogStoreSubscriber). web and mcp read the store through PluginLogStore and NodeOutputPager.

Reference rules (load on demand)

The detailed reference that used to live here is in path-scoped rules: each loads when you read, write or edit a matching file (a rule's paths are relative to the folder that holds its .claude). To read one without touching its files, open it.

TopicRuleLoads for
Node engine policy (retry, onError, trigger, timeouts, edgeBuffer), node status, node timeouts, breakpoints and tracingcore/.claude/rules/engine-node-policy.mdservice/orchestrator/**, model/workflow/**, service/control/valve/**
Restart archive, checkpoints, resume, execution log, Plus split, feature gatescore/.claude/rules/restart-and-execution-log.mdservice/archive, service/log, service/store, ExecutionResumer, config/**
Scheduler limits, idempotent starts, control audit, stop modes, gateway commands, signals and approvalscore/.claude/rules/execution-limits-and-control.mdservice/control/**, model/control/**
Definition stores and versionscore/.claude/rules/definitions-and-versions.mdservice/workflow/**, service/session/**, model/session/**
Remote workers, high availability (leader lease)core/.claude/rules/workers-and-ha.mdservice/worker/**, service/leader/**
Supervisors, ControlReviewer, the event feedcore/.claude/rules/supervisors.mdservice/supervisor/**, service/control/directive/**
Input requests, InputBroker, redactioncore/.claude/rules/input-requests.mdservice/input/**, service/redact/**
Tracker queue, execution results, validation limits, metricscore/.claude/rules/tracker-results-and-metrics.mdservice/orchestrator/tracker, .../validator, service/result, service/streaming
PostgreSQL storesplus/.claude/rules/postgres-stores.mdplus/persistence-postgres/**
SUB_WORKFLOWplugins/.claude/rules/sub-workflows.mdplugins/processors/internal/**
LLM and MCP-tool budgetsplugins/.claude/rules/llm-budgets.mdllm-call, mcp-tool-call
${env.X}, expression safety, script and process limitsplugin-api/.claude/rules/expressions-and-environment.mdplugin-api/src/main/java/**/util/**

Key File Locations

  • Services: src/main/java/com/infenia/yukta/service/
  • Validation: src/main/java/com/infenia/yukta/validation/
  • Module reference (architecture, feature catalog, settings, metrics): ../docs/contributing/core-reference.mdx. When you add, change or remove a core feature, setting or meter, update it and run core/scripts/check-core-reference.sh (then --stamp); the page's "Maintaining this document" section says what to edit.
  • Module guide (components, scenarios, ADRs): ../docs/contributing/modules/core.mdx; open risks, bugs and features: GitHub issues labelled core (gh issue list -l core); assessment and plans: ../design/assessments/core.md, ../design/plans/

Testing speed

Follow the tiered "Test Execution Policy (speed)" in the root CLAUDE.md: run --tests for the classes you changed first; full :core:test only when many files changed.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.